Question

I am attempting to modify some existing C++ code to work with my needs, but having never used C++ before, I am having some difficulties.

My goal is:

--> time and memory-intensive processes for preparation

for each file in directory:
    open file;
    generate a tagged representation; //the current code just does this
    write file; //different directory but same filename

The reason I do not want to just call the C++ program for each file (with, for instance, a shell script) is that prior to the below code running, time and memory-intensive pre-processing steps are performed. (These take about 45-60sec. while the code only takes about 2-5sec. to run.)

I have pasted the section of the code below. I want to read the arguments from the command line.

int main(int argc, char** argv) {
  /*
  pre-processing stuff
  */

  /* for each file */
  HANDLE hFind = INVALID_HANDLE_VALUE;
  string path = argv[1];
  string outpath = argv[2];
  WIN32_FIND_DATA ffd;

  //EDIT 2:
  cout << "Path: " << path << '\n'; 
  cout << "Outpath: " << outpath << '\n';

  hFind = FindFirstFile(path.c_str(), &ffd);
  if (hFind == INVALID_HANDLE_VALUE) {
    cout << "error searching directory\n";
    return false;
  }

  do {
    //istream *is(&std::cin);
    string filePath = path + ffd.cFileName;
    ifstream in( filePath.c_str() );
    if (in) {
      /* for each line */
      string line;
      int n = 1;
      string str;
      string fullOutpath = outpath + ffd.cFileName;
      ofstream File;
      File.open(fullOutpath);
      while (getline(in, line)) {
        if (line.size() > 1024) {
          cerr << "warning: the sentence seems to be too long at line " << n;
          cerr << " (please note that the input should be one-sentence-per-line)." << endl;
        }

        string postagged = bidir_postag(line, vme, vme_chunking, dont_tokenize);

        /* output to file */
        File << postagged << endl;
        //cout << postagged << endl;

        /* increment counter */
        n++;
      }
      File.close();
    } else {
      cout << "Problem opening file " << ffd.cFileName << "\n";
    }
  } while (FindNextFile(hFind, &ffd) != 0);

  if (GetLastError() != ERROR_NO_MORE_FILES) {
    cout << "Something went wrong during searching\n"; 
  }
  return true;
}

Currently, I am getting a compiler error: EDIT: compiler error fixed, thanks Blood!, but see below…

error: no matching function for call to 'std::basic_ofstream<char>::open<std::string&>

Any thoughts? Please let me know if you need more code/information. Also, I should add that I’m running these on Windows XP using command prompt.

Thanks.

EDIT:

It now compiles (thanks Blood), though when it runs it is only attempting to open the directory, not the files in the directory.

Problem opening file directory_name.

The ifstream should be opening the files in teh directory, not the directory itself.

EDIT 2:

I am running the executable fromt he command line with the following prompt:

.\tag.exe C:\indir C:\outdir

I have also tried:

.\tag.exe C:\indir\* C:\outdir\

This enumerates all the files, but how can I capture them? Also, is there a simpler way to modify my code/input?

I have also tried:

.\tag.exe C:\indir\ C:\outdir\

This gives: error searching directory.

EDIT 3:

Using:

.\tag.exe "C:\indir\*" C:\outdir\

I get the output:

Problem opening file .

Problem opening file ..

Problem opening file 2967

Problem opening file 2966

Problem opening file 4707

etc. (100s)

Solution:

Here are the key changes to the code (thanks Nate Kohl!):

string path = argv[1];
path += "\\*";

hFind = FindFirstFile(path.c_str(),&ffd);

    // in the 'do-while' loop
    string filePath = argv[1];
    filePath += "\\";
    filePath += ffd.cFileName;

    ifstream in(filePath.c_str());

    //regarding the outpath
    fullOutpath = outpath + "\\";
    fullOutpath += ffd.cFileName;
    File.open(fullOutpath.c_str());

and from the command line:

.\tag.exe C:\indir C:\outdir

The help was very much appreciated.

Answer 1

Make sure you’re passing the right path format to FindFirstFile.

From the documentation:

To examine a directory that is not a root directory, use the path to
that directory, without a trailing backslash. For example, an argument
of “C:\Windows” returns information about the directory “C:\Windows”,
not about a directory or file in “C:\Windows”. To examine the files
and directories in “C:\Windows”, use an lpFileName of “C:\Windows\*“.

---

Edit:

I’m not near a windows box right now (so this may not compile!) but I imagine that “loop over each file in a directory” would look something like this:

// argv[1] is the input path with no trailing characters, e.g. "c:\indir"

// add a wildcard because FindFirstFile expects e.g. "c:\indir\*"
TCHAR wildcard_path[MAX_PATH];
PathCombine(wildcard_path, argv[1], "*"); 

// iterate over each file
WIN32_FIND_DATA ffd;
HANDLE hFind = FindFirstFile(wildcard_path, &ffd);
if (hFind == INVALID_HANDLE_VALUE) { } // error

do {
   // ignore directories
   if (!(ffd.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)) {

      // create a full path for each file we find, e.g. "c:\indir\foo.txt"
      TCHAR file_path[MAX_PATH];
      PathCombine(file_path, argv[1], ffd.cFileName);

      // ...and do something with file_path.
   }
} while (FindNextFile(hFind, &ffd) != 0);

FindClose(hFind);